Articles

Access the latest knowledge in applied science, electrical engineering, computer science and information technology, education, and health.

Advanced Search
Filter Icon

Filters article

Years

FAQ Arrow
Single Year
Range Year
0
0

Source Title

FAQ Arrow

Authors

FAQ Arrow

29,939 Article Results

Enhancing facial recognition accuracy through feature extractions and artificial neural networks

10.11591/ijai.v14.i2.pp1056-1066
Adhi Kusnadi , Ivranza Zuhdi Pane , Fenina Adline Twince Tobing
Facial recognition is a biometric system used to identify individuals through faces. Although this technology has many advantages, it still faces several challenges. One of the main challenges is that the level of accuracy has yet to reach its maximum potential. This research aims to improve facial recognition performance by applying the discrete cosine transform (DCT) and Gaussian mixture model (GMM), which are then trained with backward propagation of errors (backpropagation) and convolutional neural networks (CNN). The research results show low DCT and GMM feature extraction accuracy with backpropagation of 4.88%. However, the combination of DCT, GMM, and CNN feature extraction produces an accuracy of up to 98.2% and a training time of 360 seconds on the Olivetti Research Laboratory (ORL) dataset, an accuracy of 98.9% and a training time of 1210 seconds on the Yale dataset, and 100% accuracy and training time 1749 seconds on the Japanese female facial expression (JAFFE) dataset. This improvement is due to the combination of DCT, GMM, and CNN's ability to remove noise and study images accurately. This research is expected to significantly contribute to overcoming accuracy challenges and increasing the flexibility of facial recognition systems in various practical situations, as well as the potential to improve security and reliability in security and biometrics.
Volume: 14
Issue: 2
Page: 1056-1066
Publish at: 2025-04-01

The 360° beach video: a supporting mindfulness intervention with virtual reality

10.11591/ijict.v14i1.pp250-258
Rohmatus Naini , Mungin Eddy Wibowo , Edy Purwanto , Mulawarman Mulawarman , E. Oos M. Anwas
This article describes optimizing virtual reality (VR) with a 360° beach video model used for mindfulness interventions. Using VR with 360° beach videos to support the presence of an immersive environment can effectively support mindfulness practices. Students are interested in the integration of technology in school counseling. VR helps in creating immersive environments such as forests, beaches, waterfalls, etc. so that students focus more on practicing mindfulness and attention in the current moment. This article focuses on optimizing 360° beach videos in the breathing mindfulness process so that it helps bring out real experiences. Obstacles to practicing mindfulness include lack of focus, mind wandering and not concentrating. through the use of 360° beach videos with VR can increase focus and be more effective in mindfulness practice.
Volume: 14
Issue: 1
Page: 250-258
Publish at: 2025-04-01

Enhancing spam detection using Harris Hawks optimization algorithm

10.12928/telkomnika.v23i2.26615
Mosleh; Al-Ahliyya Amman University M. Abualhaj , Sumaya; Al-Ahliyya Amman University Nabil Alkhatib , Ahmad; Al-Ahliyya Amman University Adel Abu-Shareha , Adeeb; Al-Ahliyya Amman University M. Alsaaidah , Mohammed; Universiti Sains Malaysia (USM) Anbar
This paper employs machine learning (ML) algorithms to identify and classify spam emails. The Harris Hawks optimization (HHO) algorithm can detect the crucial features that distinguish spam from ham emails. The HHO algorithm decreased the number of features in the ISCX-URL2016 spam dataset from 72 to 10. Implementing this will enhance the efficiency and cognitive acquisition of the ML algorithms. The decision tree (DT), Naive Bayes (NB), and AdaBoost algorithms are evaluated and contrasted to identify spam emails. The random search algorithm is used to optimize the significant hyperparameters of each algorithm for the specific task of spam identification. All three ML algorithms showed exceptional accuracy in detecting spam emails during the conducted testing. The DT algorithm attained a remarkable accuracy rate of 99.75%. The AdaBoost algorithm ranks second with an incredible accuracy of 99.67%. Finally, the NB algorithm attained an accuracy of 96.30%. The results demonstrate that the HHO algorithm shows promise in recognizing the crucial features of spam emails.
Volume: 23
Issue: 2
Page: 447-454
Publish at: 2025-04-01

Detecting fake news through deep learning: a current systematic review

10.12928/telkomnika.v23i2.26110
Idza Aisara; Universiti Malaysia Terengganu Norabid , Masita; Universiti Malaysia Terengganu Jalil , Rozniza; Universiti Malaysia Terengganu Ali , Noor Hafhizah; Universiti Malaysia Terengganu Abd Rahim
This systematic review explores the domain of deep learning-based fake new detection employing advanced search practices on Scopus and Web of Science (WoS) databases with keywords “fake news,” “deep learning,” and “method.” The study encompasses 33 articles categorized into three main themes: i) dataset and benchmarking for fake news detection, ii) multimodal approaches for fake news detection, and iii) deep learning applications and techniques for fake news detection. The analysis reveals the significance of curated datasets and robust benchmarking in improving the efficacy of fake news detection models. Additionally, the review highlights the emergence of multimodal approaches that integrate textual and visual information for improved detection accuracy. The findings clarify the essential role of deep learning applications, emphasizing the development of sophisticated models for automated identification of fake news. This systematic study adds to a thorough grasp of current research trends and offers insightful information for future developments in the field of deep learning-based false news identification.
Volume: 23
Issue: 2
Page: 329-339
Publish at: 2025-04-01

Factors influencing the integration of web accessibility in Moroccan public e-services

10.11591/ijict.v14i1.pp77-90
Chadli Fatima Ezzahra , Aniss Moumen , Driss Gretete , Zineb Sabri
Governments worldwide are increasingly digitizing their services to enhance efficiency, transparency, and accessibility for citizens. Morocco has made significant strides in adopting information and communication technology (ICT) and has implemented various initiatives to promote digital transformation across sectors. However, ensuring that digital content and e-services are accessible to everyone, including people with disabilities, is crucial to building an inclusive digital environment. Against this background, this study, based on a qualitative analysis, explores the main factors influencing the integration of web accessibility in the Moroccan public sector from the perspective of web developers and information technology (IT) managers. Through semi-structured interviews and thematic analysis, the findings reveal key barriers such as limited awareness, training deficiencies, and lack of legal framework and available guidelines. Additionally, the study highlights the need for robust managerial backing and greater collaboration with stakeholders, including people with disabilities. By raising awareness and providing actionable insights, this study offers valuable recommendations for policymakers and moves the field forward, providing a foundation for future strategies to enhance web accessibility in the Moroccan public sector.
Volume: 14
Issue: 1
Page: 77-90
Publish at: 2025-04-01

Improving visual perception through technology: a comparative analysis of real-time visual aid systems

10.12928/telkomnika.v23i2.26455
Othmane; Sidi Mohamed Ben Abdellah University Sebban , Ahmed; De Vinci Higher Education, De Vinci Research Center Azough , Mohamed; Sidi Mohamed Ben Abdellah University Lamrini
Visually impaired individuals continue to face barriers in accessing reading and listening resources. To address these challenges, we present a comparative analysis of cutting-edge technological solutions designed to assist people with visual impairments by providing relevant feedback and effective support. Our study examines various models leveraging InceptionV3 and V4 architectures, long short-term memory (LSTM) and gated recurrent unit (GRU) decoders, and datasets such as Microsoft Common Objects in Context (MSCOCO) 2017. Additionally, we explore the integration of optical character recognition (OCR), translation tools, and image detection techniques, including scale-invariant feature transform (SIFT), speeded-up robust features (SURF), oriented FAST and rotated BRIEF (ORB), and binary robust invariant scalable keypoints (BRISK). Through this analysis, we highlight the advancements and potential of assistive technologies. To assess these solutions, we have implemented a rigorous benchmarking framework evaluating accuracy, usability, response time, robustness, and generalizability. Furthermore, we investigate mobile integration strategies for real-time practical applications. As part of this effort, we have developed a mobile application incorporating features such as automatic captioning, OCR based text recognition, translation, and text-to-audio conversion, enhancing the daily experiences of visually impaired users. Our research focuses on system efficiency, user accessibility, and potential improvements, paving the way for future innovations in assistive technology.
Volume: 23
Issue: 2
Page: 249-370
Publish at: 2025-04-01

Memory management of firewall filtering rules using modified tree rule approach

10.11591/ijict.v14i1.pp141-152
Dhwani Hakani , Palvinder Singh Mann
Firewalls are essential for safety and are used for protecting a great deal of private networks. A firewall’s goal is to examine every incoming and outgoing data before granting access. A notable kind of conventional firewall is the rule-based firewall. However, when it comes to job performance, traditional listed-rule firewalls are limited, and they become useless when utilized with some networks that have extremely big firewall rule sets. This study proposes a model firewall architecture called “TreeRule Firewall,” which has benefits and functions effectively in large-scale networks like “cloud.” In order to improve cloud network security, this study suggests a modified tree rule firewall (MTRF cloud) that eliminates rule discrepancies. For the matching firewall policy, this work creates a tree rule firewall. There are no duplicate rules created by the proposed improved tree rule firewall. Also, memory utilization of different size rules is compared.
Volume: 14
Issue: 1
Page: 141-152
Publish at: 2025-04-01

Real-time age-range recognition and gender identification system through facial recognition

10.11591/ijai.v14.i2.pp992-999
Carlos Cruz-Colan , David Lopez-Herrera , Ernesto Paiva-Peredo , Kevin Acuna-Condori
Facial recognition and age estimation are being implemented in apparel retailing which is undergoing significant changes due to fashion and technology. To improve interaction with customers and refine marketing strategies. The paper proposes an approach based on a Siamese neural network and the use of tools such as MediaPipe for face detection and DeepFace for age and gender estimation. In addition, the four stages of the research work, real-time image capture, ID assignment, facial feature extraction, and data storage, are described. Early approaches to age estimation were based on biometric features, such as eyes, nose, mouth, and chin, resulting in limited accuracy and low performance in older adults. To improve accuracy, additional elements, such as the presence of wrinkles, were considered and a diverse database of images was used. The proposed methodology achieves a positive result for real-time age classification and gender ID. The results include information on gender, age, ID, time and date for each person identified.
Volume: 14
Issue: 2
Page: 992-999
Publish at: 2025-04-01

Imposing neural networks and PSO optimization in the quest for optimal ankle-foot orthosis dynamic modelling

10.12928/telkomnika.v23i2.25876
Annisa; Universiti Malaysia Sarawak Jamali , Aida Suriana; Universiti Malaysia Sarawak Abdul Razak , Shahrol; Shibaura Institute of Technology Mohamaddan
Individuals with abnormal walking patterns due to various conditions face significant challenges in daily activities, especially walking. Ankle-foot orthosis (AFO) devices are crucial in providing essential support to their lower limbs. Accurately modeling the dynamic behavior of AFO systems, particularly in predicting ground reaction forces, is a complex yet vital task to ensure their effectiveness. This research develops dynamic models for AFO systems using advanced modeling techniques, employing both parametric and non-parametric approaches. Parametric methods, such as particle swarm optimization (PSO), and non-parametric methods, like multi-layer perceptron (MLP) neural networks, are utilized through system identification methods. According to the findings, the MLP neural network continuously generates objective results and performs exceptionally well in correctly detecting the AFO system, attaining a noticeably lower mean squared prediction error of 0.000011. This research highlights the potential of advanced modeling techniques, particularly MLP neural networks, in enhancing AFO system modeling accuracy. Although parametric techniques like PSO are useful, the MLP approach performs better, offering insightful information about modelling AFO systems and indicating that non-parametric techniques like MLP neural networks have potential to further AFO creation and control.
Volume: 23
Issue: 2
Page: 484-494
Publish at: 2025-04-01

Symmetrical cryptographic algorithms in the lightweight internet of things

10.11591/ijict.v14i1.pp307-314
Akshaya Dhingra , Vikas Sindhu , Anil Sangwan
The internet of things (IoT) has emerged as a prominent area of scrutiny. It is being deployed in multiple applications like smart homes, smart agriculture, intelligent surveillance systems, and even innovative industries. Security is a significant issue that needs to be addressed in these types of networks. This paper aims to describe symmetrical lightweight cryptographic algorithms (SLCAs) for lightweight IoT networks. The article focuses on discussing the principal difficulties of using cryptography in lightweight IoT devices, exploring SLCAs and their types based on structure formation throughout the literature survey, and comparing and evaluating different LCAs proposed in recent research. The main goal is to demonstrate how to solve the issues associated with conventional cryptography techniques and how lightweight cryptography algorithms aid limited IoT devices in achieving cybersecurity objectives.
Volume: 14
Issue: 1
Page: 307-314
Publish at: 2025-04-01

An ensemble image augmentation approach to enhance granular parakeratosis dataset

10.11591/ijeecs.v38.i1.pp312-320
Sheetal Janthakal , Girisha Hosalli
The study discusses the revolutionizing impact of deep convolutional neural network (CNN) techniques on medical image classification, particularly in identifying skin lesions. It addresses the challenge of limited datasets for granular parakeratosis (GP) and paraneoplastic pemphigus (PNP) by employing traditional and advanced ensemble data augmentation techniques. These techniques include geometric transformations, generative adversarial networks (GANs), Cutout, and keep augment. GP affects keratinization in the groin and other regions, while PNP is associated with malignancies. The study’s relevance is enhanced by the shared imaging characteristics of the chosen conditions. By utilizing tools like U-net for segmentation, region props for feature extraction, and a support vector machine (SVM) 10-fold cross-validation model for classification, the study achieved impressive performance metrics, including 95% accuracy, 100% sensitivity, and 100% specificity when evaluated on the DermnetNZ skin lesion dataset. These findings underscore the effectiveness of augmentation in enhancing the precision of medical image classifiers and signify a substantial improvement over traditional method. Thus, the research showcases the critical role of data augmentation in overcoming data scarcity challenges and advances medical image analysis.
Volume: 38
Issue: 1
Page: 312-320
Publish at: 2025-04-01

Advanced crop yield prediction using machine learning and deep learning: a comprehensive review

10.12928/telkomnika.v23i2.26621
Ayush; Manipal University Jaipur Anand , Kavita; Manipal University Jaipur Jhajharia
The advancement of machine learning (ML) and deep learning (DL) techniques has significantly improved crop yield prediction, making it more accurate and reliable. In this review, the implementation of ML and DL algorithms for crop yield prediction is thoroughly investigated, focusing on their crucial role in enhancing crop productivity. Along with ML and DL algorithms examine, the review analyses the use of remote sensing technologies, such as satellite and drone data, in providing high-resolution inputs essential for accurate yield predictions. The study identifies the state of art algorithms, most used features, data sources and evaluation metrics, providing a comparison of ML and DL. The findings indicate that DL models are more effective with large datasets, while ML models remain robust for smaller datasets. The future directions are proposed to develop the generalised models for different crops and regions. The review aims to assist researchers by summarising state of art techniques and identifying the present.
Volume: 23
Issue: 2
Page: 402-415
Publish at: 2025-04-01

Analyzing the impact of sports activity intensity on muscle capacity through integrated biosensor technology

10.12928/telkomnika.v23i2.26264
Ervin; Bandung State Polytechnic Masita Dewi , Nurista; Bandung State Polytechnic Wahyu Kirana , Sugondo; Telkom University Hadiyoso
In the past few years, biosensor technology has paved the way for new insights into the physiological effects of physical exercise. Quantitative analysis, especially in the case of muscle capacity measurement, is the focus of studies to assess the impact of sports activities. Therefore, this study examines the impact of sports activity intensity on muscle capacity using an integrated biosensor system developed at Bandung State Polytechnic. Surface electromyography (sEMG) measurements were conducted on 30 participants aged 20–25 during various sports activities. Results showed a strong positive correlation (r=0.814) between sports activity frequency and muscle contraction, suggesting higher activity correlates with increased muscle activity. Conversely, the correlation during muscle relaxation was low (r=0.261), indicating independence from sports activity. In the future, it is expected that integrated biosensors will have the ability to concurrently measure and monitor various parameters like heart rate (via electrocardiogram), blood oxygen levels (via photoplethysmography), and blood pressure. The integrated biosensor system allows for comprehensive assessment and optimization of sports performance and injury prevention strategies.
Volume: 23
Issue: 2
Page: 466-472
Publish at: 2025-04-01

Enhancing fall detection and classification using Jarratt‐butterfly optimization algorithm with deep learning

10.11591/ijai.v14.i2.pp1461-1470
Kakirala Durga Bhavani , Melkias Ferni Ukrit
Falls pose significant risk to the health and safety of individuals, specifically for vulnerable populations as the elderly and those with specific medical conditions. The repercussions of falls can be severe, leading to injuries, loss of independence, and increased healthcare costs. Consequently, the development of effective fall detection systems is crucial for providing timely assistance and enhancing the overall well-being of affected individuals. Recent advancements in deep learning (DL) have opened new avenues for automating fall detection through the analysis of sensor data and video footage. DL algorithms are especially well-suited for this task because they can automatically learn complex features and patterns from raw data, eliminating the need for extensive manual feature engineering. This article introduces a novel approach to fall detection and classification, termed the fall detection and classification using Jarratt‐butterfly optimization algorithm with deep learning (FDC-JBOADL) algorithm. The FDC-JBOADL technique employs a median filtering (MF) method to mitigate noise and utilizes the EfficientNet model for robust feature extraction, capturing both motion patterns and appearance characteristics of individuals. Furthermore, the classification of fall events is achieved through a long short-term memory (LSTM) classifier, with hyperparameter optimization facilitated by Jarratt‐butterfly optimization algorithm (JBOA). Through a comprehensive series of experiments, the efficacy of FDC-JBOADL technique is validated, demonstrating superior performance compared to existing methodologies in the domain of fall detection.
Volume: 14
Issue: 2
Page: 1461-1470
Publish at: 2025-04-01

A hybrid approach of pattern recognition to detect marine animals

10.11591/ijict.v14i1.pp240-249
Vijayalakshmi Balachandran , Thanga Ramya Shanmugavel , Ramar Kadarkarayandi , Vijayalakshmi Kandhasamy
Acquiring up-to-date knowledge about various animals will have a significant impact on effectively managing species within the ecosystem. Manually identifying animals and their traits continues to be a costly and time-consuming process. The development of a system using the most recent developments in computer vision machine learning was necessary to address the issues of detecting sharks and aquatic species in areas filled with surfers, rocks, and various other potential false positives. In the ocean most of the species are cold-blooded animals hence they cannot be tracked with thermal cameras. Ocean’s dynamic environment affects simple techniques like color separation, intensity histograms, and optical flow. Hence a hybrid approach using convolutional neural network - support vector machine (CNN-SVM) classifier is proposed to perform the pattern recognition. A CNN is employed for feature extraction by using the histogram of gradients value. Subsequently, a SVM classifier is employed to identify and categorise marine species in the vicinity of the seacoast. This serves to notify individuals who engage in swimming activities in the ocean. The suggested model is evaluated against alternative machine learning approaches, and it achieves a superior accuracy of 95% compared to the others.
Volume: 14
Issue: 1
Page: 240-249
Publish at: 2025-04-01
1 2 3
...
1995 1996
Show 232 of 1996

Discover Our Library

Embark on a journey through our expansive collection of articles and let curiosity lead your path to innovation.

Explore Now
Library 3D Ilustration